Re: Hot Standby Design

On Wed, 2008-09-24 at 08:28 +0100, Dave Page wrote:
> On Wed, Sep 24, 2008 at 5:30 AM, Robert Treat
> <xzilla(at)users(dot)sourceforge(dot)net> wrote:
>
> > I presume this means the backend kill function would work? Also, can you go
> > into why pgAgent would not work? (I presume it's because you can't update
> > information that needs to be changed when jobs run, if thats the case it
> > might be worth thinking about making pgAgent work across different clusters)
>
> Probably - it needs to be able to update its catalogs to record job
> results and ensure that only a single node runs any given job
> instance.
>
> We could probably update the code to optionally accept a connection
> string instead of a database name, which would allow us to operate
> against servers other than the one with the catalogs installed.

It would be useful to be able to schedule jobs against multiple nodes.

Not much need for maintenance actions on a Standby node, but I think if
I say "there aren't any needed at all" that would likely be wrong.

--
Simon Riggs www.2ndQuadrant.com
PostgreSQL Training, Services and Support